No Cover Image

E-Thesis 469 views 271 downloads

Feature Driven Learning Techniques for 3D Shape Segmentation / David George

Swansea University Author: David George

DOI (Published version): 10.23889/Suthesis.51429

Abstract

Segmentation is a fundamental problem in 3D shape analysis and machine learning. The abil-ity to partition a 3D shape into meaningful or functional parts is a vital ingredient of many down stream applications like shape matching, classification and retrieval. Early segmentation methods were based on...

Full description

Published: Swansea 2019
Institution: Swansea University
Degree level: Doctoral
Degree name: Ph.D
Supervisor: Tam, Gary K.L. ; Xie Xianghua
URI: https://cronfa.swan.ac.uk/Record/cronfa51429
Tags: Add Tag
No Tags, Be the first to tag this record!
first_indexed 2019-08-15T21:29:50Z
last_indexed 2021-04-10T03:15:36Z
id cronfa51429
recordtype RisThesis
fullrecord <?xml version="1.0"?><rfc1807><datestamp>2021-04-09T17:04:28.0569515</datestamp><bib-version>v2</bib-version><id>51429</id><entry>2019-08-15</entry><title>Feature Driven Learning Techniques for 3D Shape Segmentation</title><swanseaauthors><author><sid>589852950a34dca85d753df9ebdb8e14</sid><ORCID>NULL</ORCID><firstname>David</firstname><surname>George</surname><name>David George</name><active>true</active><ethesisStudent>true</ethesisStudent></author></swanseaauthors><date>2019-08-15</date><abstract>Segmentation is a fundamental problem in 3D shape analysis and machine learning. The abil-ity to partition a 3D shape into meaningful or functional parts is a vital ingredient of many down stream applications like shape matching, classification and retrieval. Early segmentation methods were based on approaches like fitting primitive shapes to parts or extracting segmen-tations from feature points. However, such methods had limited success on shapes with more complex geometry. Observing this, research began using geometric features to aid the segmen-tation, as certain features (e.g. Shape Diameter Function (SDF)) are less sensitive to complex geometry. This trend was also incorporated in the shift to set-wide segmentations, called co-segmentation, which provides a consistent segmentation throughout a shape dataset, meaning similar parts have the same segment identifier. The idea of co-segmentation is that a set of same class shapes (i.e. chairs) contain more information about the class than a single shape would, which could lead to an overall improvement to the segmentation of the individual shapes. Over the past decade many different approaches of co-segmentation have been explored covering supervised, unsupervised and even user-driven active learning. In each of the areas, there has been widely adopted use of geometric features to aid proposed segmentation algorithms, with each method typically using different combinations of features. The aim of this thesis is to ex-plore these different areas of 3D shape segmentation, perform an analysis of the effectiveness of geometric features in these areas and tackle core issues that currently exist in the literature.Initially, we explore the area of unsupervised segmentation, specifically looking at co-segmentation, and perform an analysis of several different geometric features. Our analysis is intended to compare the different features in a single unsupervised pipeline to evaluate their usefulness and determine their strengths and weaknesses. Our analysis also includes several features that have not yet been explored in unsupervised segmentation but have been shown effective in other areas.Later, with the ever increasing popularity of deep learning, we explore the area of super-vised segmentation and investigate the current state of Neural Network (NN) driven techniques. We specifically observe limitations in the current state-of-the-art and propose a novel Convolu-tional Neural Network (CNN) based method which operates on multi-scale geometric features to gain more information about the shapes being segmented. We also perform an evaluation of several different supervised segmentation methods using the same input features, but with vary-ing complexity of model design. This is intended to see if the more complex models provide a significant performance increase.Lastly, we explore the user-driven area of active learning, to tackle the large amounts of inconsistencies in current ground truth segmentation, which are vital for most segmentation methods. Active learning has been used to great effect for ground truth generation in the past, so we present a novel active learning framework using deep learning and geometric features to assist the user in co-segmentation of a dataset. Our method emphasises segmentation accu-racy while minimising user effort, providing an interactive visualisation for co-segmentation analysis and the application of automated optimisation tools.In this thesis we explore the effectiveness of different geometric features across varying segmentation tasks, providing an in-depth analysis and comparison of state-of-the-art methods.</abstract><type>E-Thesis</type><journal/><volume/><journalNumber/><paginationStart/><paginationEnd/><publisher/><placeOfPublication>Swansea</placeOfPublication><isbnPrint/><isbnElectronic/><issnPrint/><issnElectronic/><keywords>3D segmentation, shape analysis, machine learning</keywords><publishedDay>31</publishedDay><publishedMonth>12</publishedMonth><publishedYear>2019</publishedYear><publishedDate>2019-12-31</publishedDate><doi>10.23889/Suthesis.51429</doi><url/><notes/><college>COLLEGE NANME</college><department>Computer Science</department><CollegeCode>COLLEGE CODE</CollegeCode><institution>Swansea University</institution><supervisor>Tam, Gary K.L. ; Xie Xianghua</supervisor><degreelevel>Doctoral</degreelevel><degreename>Ph.D</degreename><apcterm/><lastEdited>2021-04-09T17:04:28.0569515</lastEdited><Created>2019-08-15T18:34:22.1647579</Created><path><level id="1">Faculty of Science and Engineering</level><level id="2">School of Mathematics and Computer Science - Computer Science</level></path><authors><author><firstname>David</firstname><surname>George</surname><orcid>NULL</orcid><order>1</order></author></authors><documents><document><filename>51429__15009__ecd0e3db1f47499cb460994c78c6595b.pdf</filename><originalFilename>George_David_PhD_Thesis_Final.pdf</originalFilename><uploaded>2019-08-15T18:35:06.1670000</uploaded><type>Output</type><contentLength>22174727</contentLength><contentType>application/pdf</contentType><version>E-Thesis &#x2013; open access</version><cronfaStatus>true</cronfaStatus><embargoDate>2019-08-15T00:00:00.0000000</embargoDate><documentNotes>Copyright: The author, David George, 2019.</documentNotes><copyrightCorrect>true</copyrightCorrect><language>eng</language></document></documents><OutputDurs/></rfc1807>
spelling 2021-04-09T17:04:28.0569515 v2 51429 2019-08-15 Feature Driven Learning Techniques for 3D Shape Segmentation 589852950a34dca85d753df9ebdb8e14 NULL David George David George true true 2019-08-15 Segmentation is a fundamental problem in 3D shape analysis and machine learning. The abil-ity to partition a 3D shape into meaningful or functional parts is a vital ingredient of many down stream applications like shape matching, classification and retrieval. Early segmentation methods were based on approaches like fitting primitive shapes to parts or extracting segmen-tations from feature points. However, such methods had limited success on shapes with more complex geometry. Observing this, research began using geometric features to aid the segmen-tation, as certain features (e.g. Shape Diameter Function (SDF)) are less sensitive to complex geometry. This trend was also incorporated in the shift to set-wide segmentations, called co-segmentation, which provides a consistent segmentation throughout a shape dataset, meaning similar parts have the same segment identifier. The idea of co-segmentation is that a set of same class shapes (i.e. chairs) contain more information about the class than a single shape would, which could lead to an overall improvement to the segmentation of the individual shapes. Over the past decade many different approaches of co-segmentation have been explored covering supervised, unsupervised and even user-driven active learning. In each of the areas, there has been widely adopted use of geometric features to aid proposed segmentation algorithms, with each method typically using different combinations of features. The aim of this thesis is to ex-plore these different areas of 3D shape segmentation, perform an analysis of the effectiveness of geometric features in these areas and tackle core issues that currently exist in the literature.Initially, we explore the area of unsupervised segmentation, specifically looking at co-segmentation, and perform an analysis of several different geometric features. Our analysis is intended to compare the different features in a single unsupervised pipeline to evaluate their usefulness and determine their strengths and weaknesses. Our analysis also includes several features that have not yet been explored in unsupervised segmentation but have been shown effective in other areas.Later, with the ever increasing popularity of deep learning, we explore the area of super-vised segmentation and investigate the current state of Neural Network (NN) driven techniques. We specifically observe limitations in the current state-of-the-art and propose a novel Convolu-tional Neural Network (CNN) based method which operates on multi-scale geometric features to gain more information about the shapes being segmented. We also perform an evaluation of several different supervised segmentation methods using the same input features, but with vary-ing complexity of model design. This is intended to see if the more complex models provide a significant performance increase.Lastly, we explore the user-driven area of active learning, to tackle the large amounts of inconsistencies in current ground truth segmentation, which are vital for most segmentation methods. Active learning has been used to great effect for ground truth generation in the past, so we present a novel active learning framework using deep learning and geometric features to assist the user in co-segmentation of a dataset. Our method emphasises segmentation accu-racy while minimising user effort, providing an interactive visualisation for co-segmentation analysis and the application of automated optimisation tools.In this thesis we explore the effectiveness of different geometric features across varying segmentation tasks, providing an in-depth analysis and comparison of state-of-the-art methods. E-Thesis Swansea 3D segmentation, shape analysis, machine learning 31 12 2019 2019-12-31 10.23889/Suthesis.51429 COLLEGE NANME Computer Science COLLEGE CODE Swansea University Tam, Gary K.L. ; Xie Xianghua Doctoral Ph.D 2021-04-09T17:04:28.0569515 2019-08-15T18:34:22.1647579 Faculty of Science and Engineering School of Mathematics and Computer Science - Computer Science David George NULL 1 51429__15009__ecd0e3db1f47499cb460994c78c6595b.pdf George_David_PhD_Thesis_Final.pdf 2019-08-15T18:35:06.1670000 Output 22174727 application/pdf E-Thesis – open access true 2019-08-15T00:00:00.0000000 Copyright: The author, David George, 2019. true eng
title Feature Driven Learning Techniques for 3D Shape Segmentation
spellingShingle Feature Driven Learning Techniques for 3D Shape Segmentation
David George
title_short Feature Driven Learning Techniques for 3D Shape Segmentation
title_full Feature Driven Learning Techniques for 3D Shape Segmentation
title_fullStr Feature Driven Learning Techniques for 3D Shape Segmentation
title_full_unstemmed Feature Driven Learning Techniques for 3D Shape Segmentation
title_sort Feature Driven Learning Techniques for 3D Shape Segmentation
author_id_str_mv 589852950a34dca85d753df9ebdb8e14
author_id_fullname_str_mv 589852950a34dca85d753df9ebdb8e14_***_David George
author David George
author2 David George
format E-Thesis
publishDate 2019
institution Swansea University
doi_str_mv 10.23889/Suthesis.51429
college_str Faculty of Science and Engineering
hierarchytype
hierarchy_top_id facultyofscienceandengineering
hierarchy_top_title Faculty of Science and Engineering
hierarchy_parent_id facultyofscienceandengineering
hierarchy_parent_title Faculty of Science and Engineering
department_str School of Mathematics and Computer Science - Computer Science{{{_:::_}}}Faculty of Science and Engineering{{{_:::_}}}School of Mathematics and Computer Science - Computer Science
document_store_str 1
active_str 0
description Segmentation is a fundamental problem in 3D shape analysis and machine learning. The abil-ity to partition a 3D shape into meaningful or functional parts is a vital ingredient of many down stream applications like shape matching, classification and retrieval. Early segmentation methods were based on approaches like fitting primitive shapes to parts or extracting segmen-tations from feature points. However, such methods had limited success on shapes with more complex geometry. Observing this, research began using geometric features to aid the segmen-tation, as certain features (e.g. Shape Diameter Function (SDF)) are less sensitive to complex geometry. This trend was also incorporated in the shift to set-wide segmentations, called co-segmentation, which provides a consistent segmentation throughout a shape dataset, meaning similar parts have the same segment identifier. The idea of co-segmentation is that a set of same class shapes (i.e. chairs) contain more information about the class than a single shape would, which could lead to an overall improvement to the segmentation of the individual shapes. Over the past decade many different approaches of co-segmentation have been explored covering supervised, unsupervised and even user-driven active learning. In each of the areas, there has been widely adopted use of geometric features to aid proposed segmentation algorithms, with each method typically using different combinations of features. The aim of this thesis is to ex-plore these different areas of 3D shape segmentation, perform an analysis of the effectiveness of geometric features in these areas and tackle core issues that currently exist in the literature.Initially, we explore the area of unsupervised segmentation, specifically looking at co-segmentation, and perform an analysis of several different geometric features. Our analysis is intended to compare the different features in a single unsupervised pipeline to evaluate their usefulness and determine their strengths and weaknesses. Our analysis also includes several features that have not yet been explored in unsupervised segmentation but have been shown effective in other areas.Later, with the ever increasing popularity of deep learning, we explore the area of super-vised segmentation and investigate the current state of Neural Network (NN) driven techniques. We specifically observe limitations in the current state-of-the-art and propose a novel Convolu-tional Neural Network (CNN) based method which operates on multi-scale geometric features to gain more information about the shapes being segmented. We also perform an evaluation of several different supervised segmentation methods using the same input features, but with vary-ing complexity of model design. This is intended to see if the more complex models provide a significant performance increase.Lastly, we explore the user-driven area of active learning, to tackle the large amounts of inconsistencies in current ground truth segmentation, which are vital for most segmentation methods. Active learning has been used to great effect for ground truth generation in the past, so we present a novel active learning framework using deep learning and geometric features to assist the user in co-segmentation of a dataset. Our method emphasises segmentation accu-racy while minimising user effort, providing an interactive visualisation for co-segmentation analysis and the application of automated optimisation tools.In this thesis we explore the effectiveness of different geometric features across varying segmentation tasks, providing an in-depth analysis and comparison of state-of-the-art methods.
published_date 2019-12-31T04:03:18Z
_version_ 1763753274538721280
score 11.013575

Similar Items